SUMO community discussions

[important announcement] New contributor policy around the use of generative AI tools

  • 7 Replies
  • Last reply by AliceWyman
  1. Kiki
    Kiki 1280 posts
    more options

    Hey folks,

    As of today, we have a new contributor policy around the use of generative AI tools. The full policy can be seen from the following article Contributor Policy: Usage of Generative AI (e.g., ChatGPT).

    TL;DR: You may use generative AI tools in Mozilla Support, as long as you abide to our policy. The consequences are still to be determined. I'm going to discuss it further with the moderator team around the enforcement and what the right consequences will be for those who failed to adhere to the policy. Please stay tuned!

    In the meantime, please familiarize yourself with the new policy and please raise concerns, feedback, or questions in this thread.

    Last but not least, I'd like to give credit to Sarto Jama who started this initiative when I was out for maternity leave. Thanks for your work, Sarto!

    Thanks!

    Hey folks, As of today, we have a new contributor policy around the use of generative AI tools. The full policy can be seen from the following article [/kb/contributor-policy-generative-ai-usage Contributor Policy: Usage of Generative AI (e.g., ChatGPT)]. '''TL;DR: You may use generative AI tools in Mozilla Support, as long as you abide to our policy.''' The consequences are still to be determined. I'm going to discuss it further with the moderator team around the enforcement and what the right consequences will be for those who failed to adhere to the policy. Please stay tuned! In the meantime, please familiarize yourself with the new policy and please raise concerns, feedback, or questions in this thread. Last but not least, I'd like to give credit to Sarto Jama who started this initiative when I was out for maternity leave. Thanks for your work, Sarto! Thanks!
  2. TyDraniu
    TyDraniu 153 posts
    more options

    Good idea.

    I'd emphasize the need of a validation of the provided solution and/or troubleshooting steps in the latest version of said product.

    Thanks Sarto & Kiki!

    Good idea. I'd emphasize the need of a validation of the provided solution and/or troubleshooting steps in the latest version of said product. Thanks Sarto & Kiki!
  3. AliceWyman
    AliceWyman 6154 posts
    more options

    Small point: The title for this new article, Contributor Policy: Usage of Generative AI (e.g., ChatGPT) includes a colon, which prevents wikilinks. [[Contributor Policy:Usage of Generative AI (e.g., ChatGPT)]] doesn't work. This is Bug 749835 - Can't make a wiki link to an article with a colon in the title.

    If you change the colon to a hyphen, the wikilink will work.

    Small point: The title for this new article, Contributor Policy: Usage of Generative AI (e.g., ChatGPT) includes a colon, which prevents wikilinks. [[Contributor Policy: Usage of Generative AI (e.g., ChatGPT)]] doesn't work. This is [https://bugzilla.mozilla.org/show_bug.cgi?id=749835 Bug 749835] - Can't make a wiki link to an article with a colon in the title. If you change the colon to a hyphen, the wikilink will work.
  4. cor-el
    cor-el 1439 posts
    more options

    About this line:

    • Take time to validate the responses that you receive from the AI tools.

    AI chat bots are usually trained with data from a few years ago, so answers they generate may not apply to the current release. Maybe add something like:

    • , make sure they apply to the Firefox release used by the thread creator.

    Also maybe add to a rule to add a link to relevant support article(s) if possible as a chat bot reply only contains plain text and no references that can be checked.

    About this line: *Take time to validate the responses that you receive from the AI tools. AI chat bots are usually trained with data from a few years ago, so answers they generate may not apply to the current release. Maybe add something like: * '', make sure they apply to the Firefox release used by the thread creator.'' Also maybe add to a rule to add a link to relevant support article(s) if possible as a chat bot reply only contains plain text and no references that can be checked.
  5. Paul
    Paul 1083 posts
    more options

    Is there any update on the enforcement element of this? The moderator team have been looking into this for quite a few months, but it would be good to get something final in place soon.

    Is there any update on the enforcement element of this? The moderator team have been looking into this for quite a few months, but it would be good to get something final in place soon.
  6. jscher2000 - Support Volunteer
    jscher2000 - Support Volunteer 794 posts
    more options

    I just looked at this today and I can't find a reference to accuracy, correctness, or relevance. Is that because it is covered in a different article that we could link to? If not, maybe at least something along these lines:

    • Accuracy and Relevance

    Make sure to check and correct the suggested steps. Models are trained on materials for older releases of Firefox, and menu names, Settings page layout, and features may have changed. References to keyboard shortcuts should be adapted to the user's OS (for example, Linux or Mac).

    I just looked at this today and I can't find a reference to accuracy, correctness, or relevance. Is that because it is covered in a different article that we could link to? If not, maybe at least something along these lines: * '''Accuracy and Relevance''' Make sure to check and correct the suggested steps. Models are trained on materials for older releases of Firefox, and menu names, Settings page layout, and features may have changed. References to keyboard shortcuts should be adapted to the user's OS (for example, Linux or Mac).
  7. Kiki
    Kiki 1280 posts
    more options

    Hi folks,

    Thanks for sharing your feedback here. That's been really helpful for me to think about the enforcement part of this new policy.

    I'm currently discussing the topic about enforcement and consequences of this new policy with the forum moderator team. Please stay tuned for more updates!

    Hi folks, Thanks for sharing your feedback here. That's been really helpful for me to think about the enforcement part of this new policy. I'm currently discussing the topic about enforcement and consequences of this new policy with the forum moderator team. Please stay tuned for more updates!
  8. AliceWyman
    AliceWyman 6154 posts
    more options

    AliceWyman said

    Small point: The title for this new article, Contributor Policy: Usage of Generative AI (e.g., ChatGPT) includes a colon, which prevents wikilinks. [[Contributor Policy:Usage of Generative AI (e.g., ChatGPT)]] doesn't work. This is Bug 749835 - Can't make a wiki link to an article with a colon in the title. If you change the colon to a hyphen, the wikilink will work.

    (For the record) I updated the article title to Contributor Policy - Usage of Generative AI (e.g., ChatGPT) (the wikilink Contributor Policy - Usage of Generative AI (e.g., ChatGPT) now works).

    ''AliceWyman [[#post-86618|said]]'' <blockquote> Small point: The title for this new article, Contributor Policy: Usage of Generative AI (e.g., ChatGPT) includes a colon, which prevents wikilinks. [[Contributor Policy: Usage of Generative AI (e.g., ChatGPT)]] doesn't work. This is [https://bugzilla.mozilla.org/show_bug.cgi?id=749835 Bug 749835] - Can't make a wiki link to an article with a colon in the title. If you change the colon to a hyphen, the wikilink will work. </blockquote> (For the record) I updated the article title to '''Contributor Policy - Usage of Generative AI (e.g., ChatGPT)''' (the wikilink [[Contributor Policy - Usage of Generative AI (e.g., ChatGPT)]] now works).